Chlorine gas is a potential weapon of mass destruction and is one of the most commonly used chemicals [1-5]. Chlorination accidents are common, exposing thousands of people collectively each year [1, 2, 5-16]. A large accident releasing chlorine gas could result in a public health disaster [6, 17]. The acute health effects on the lungs are well documented [1, 7, 9, 18-22], but there has been no large systematic long-term longitudinal cohort study of a population exposed to a single high-concentration chlorine gas event. New science has provided evidence for potential long-term health effects [24], and called for long-term longitudinal study of irritant gas exposed persons for at least 5 years [27]. On January 6, 2005 a train derailment and subsequent 54,422 kg chlorine gas spill occurred in the cotton mill town of Graniteville, South Carolina [28-33]. Several hundred people became immediately sick and thousands were potentially exposed, including millworkers. Millworkers had their lung health monitored for byssinosis from 1980 to 2006. We have a cohort of 8,234 millworkers with at least 3 years of pre-event spirometry data. Millworkers who returned to work after the accident had lung function loss (FEV1) in both 2005 and 2006 that was twice the rate observed in the years immediately before the accident. We will assess differential pre/post-event lung function decline in a sample of millworkers with 3 years of new lung function tests contrasted with data from 3 years pre-event in a natural experiment, using a double-blinded interrupted time-series design nested within a prospective longitudinal cohort study. Hypothesis are: (H1) Lung function decline increases with increasing chlorine gas exposure (H2) Airway responsiveness increases with increasing chlorine gas exposure (H3) Age, sex, race, smoking, post traumatic stress disorder (PTSD), atopy, occupational history of lung-irritant exposures, pre-existing lung disease, and obesity modify the association of exposure to chlorine gas with lung function decline (H4) Increased lung function decline is associated with increased fibrosis (restrictive lung defects) and airway inflammation (obstructive lung defects). Assuming a 5 percent annual attrition, we will enroll 670 millworkers to study 600, including 150 millworkers from each of the 3 validated plume model exposure estimate groups (peak sustained 30-minute concentrations of 0.5 ppm, 2.8 ppm, and 28 ppm chlorine gas outdoors) and 150 unexposed millworkers. We will measure lung function, airway inflammation, indicators of airway fibrosis, atopy, PTSD, and airway responsiveness at 6-10 years post- event. This study is significant because it is the largest general US population ever exposed to a chlorine gas event. Our investigators are strong and capable, having succeeded with public health screening, validated plume modeling, and experience with other disasters. This study has the unique opportunity for strong scientific influence due to the innovative and robust study design, having personal exposure group double blinded, each participant serving as their own control, and confounder data available from before the exposure. PUBLIC HEALTH RELEVANCE: Chlorine gas is a potential weapon of mass destruction and is one of the most commonly used chemicals in industry [1-5]. In a natural experiment, using a double-blinded interrupted time-series design nested within a prospective longitudinal cohort study, we will assess differential pre/post-event lung function decline with 3 years of new post-event lung function test data contrasted with data from 3 years pre-event in a sample of 600 millworkers from our cohort of 8,234 with at least 3 years of pre-event data who were potentially exposed to chlorine gas from the January 6, 2005 chlorine gas accident in Graniteville, South Carolina. This study of the largest general US population ever exposed to a chlorine gas event has the unique opportunity for significant scientific influence due to the innovative and robust study design, having personal exposure group double blinded, each participant serving as their own control, and confounder data available from before the exposure.